SandBase Harness — Watchlist

MCP100 Assessment

Watchlist / Tracking

Why we’re tracking it

Follow harness adapters and execution controls as SandBase develops

Reviewed Oct 5, 2026 · Documentation review; complete setup workflow not independently tested

Best fit

Self-hosting managed coding-agent sessions behind a six-tool MCP bridge

What we know

Self-hosted agent harness with durable sessions and an MCP bridge for listing agents, creating/running/stopping sessions and retrieving artifacts.

Maintenance / activity evidence

Observed Oct 5, 2026 (UTC)fix(strategy): trim stale tool outputs when a turn outgrows the conte…

Tracking evidence observed Oct 5, 2026. These dated signals do not establish current maintenance. Reviewed enrichment confidence: high.

Activity source · commits ↗Activity source · releases ↗

Verification gaps / watch-outs

Installation and comprehensive security have not been independently audited.
Agent/session management authority of the Harness runtime. Writes: Create/run/stop agent sessions.
Optional Harness API key.
Sensitive access: Agent workspaces, Artifacts, API key

The MCP bridge can launch and control autonomous coding-agent sessions; protect the Harness API and workspace boundaries.

Security source · repository notice ↗Security source · policy file ↗

Full authority, access and evidence limits are available in Review context and Access & risk.

Popularity

672 repository stars

Stars observed Sep 25, 2026.

Preserve any already-reviewed adoption metrics in MCP100. Batch 4 did not refresh package downloads because the package-stats provider had insufficient credits.

All limitations / Review context

Which MCP tools and runtime controls are documented by the current harness?

The bounded assessment reviewed the official README, repository notices, security policy path, and security page for known material public-tracking blockers. No such blocker was identified. This is a public-tracking assessment, not an installation safety certification.

The MCP bridge can launch and control autonomous coding-agent sessions; protect the Harness API and workspace boundaries.
Agent/session management authority of the Harness runtime.
Optional Harness API key.
Session prompts, artifacts and workspace data are managed by the self-hosted API.
Reviewed write actions: Create/run/stop agent sessions
Sensitive access: Agent workspaces, Artifacts, API key
Installation, client compatibility and comprehensive security have not been independently audited. Authentication and access depend on the configured services.
No independent runtime, installation, or comprehensive security audit was performed.
No installation, runtime, client matrix, or complete tool-contract audit was performed in this review.
MCP bridge requires a running Harness API.
Do not bind unrelated npm packages such as `managed-agents` to this identity.

Original activity observations

2026-10-05T07:00:20Z: fix(strategy): trim stale tool outputs when a turn outgrows the conte…

Technical identity

github.com/sandbaseai/sandbase-harness

Exact implementation source: https://github.com/sandbaseai/sandbase-harness ↗
Reviewed resource type: Host
Repository identity: sandbaseai/sandbase-harness

Evidence & Sources

Evidence basis: Recorded source evidence
Setup runtime verification: Not independently runtime-tested by MCP100 for this setup workflow.

Reviewed Oct 5, 2026 · Latest source observation Oct 6, 2026

GitHub repository ↗
sandbaseai/sandbase-harness
Recorded repository material and project history.
Includes references whose attribution is not established.

Setup · Watchlist

Host / Desktop app setup

Publisher instructions.

Source evidence observed Oct 14, 2026.

Additional setup details required — see the specific gaps below.

Download / install
Publisher instructions
SandBase Harness is a full agent host/framework rather than an ordinary MCP server.
publisher setup entry · Publisher instructions
git clone --branch v0.3.8 --depth 1 https://github.com/sandbaseai/sandbase-harness.git cd sandbase-harness npm ci npm run build
Configure application / integrations
MCP100 setup guide
Requires Node 22+, npm 10+ and model provider key. From a separate workspace run node ../sandbase-harness/dist/index.js init then start. Optional Docker sandbox.
Check connection
Suggested check
Run harness CLI help/config check; do not start autonomous external actions.
Disconnect / uninstall / revoke
Suggested check
Stop the Sandbase Harness runtime, remove its configured service/container entry, and uninstall the checkout only if desired. Do not use the unrelated npm package named `managed-agents`; publisher explicitly says it is not theirs.
Client unregistration, program uninstall, persistent-data deletion, and credential revocation must be treated separately.

Official Documentation & Sources

What it does

Self-hosted agent harness with durable sessions and an MCP bridge for listing agents, creating/running/stopping sessions and retrieving artifacts.

Capability: AI & Reasoning
Resource type: Host
Industry: Software & Technology
Best fit: Self-hosting managed coding-agent sessions behind a six-tool MCP bridge

Representative uses

List agents
Create/run durable sessions
Inspect session/artifacts
Stop sessions

Limitations

MCP bridge requires a running Harness API.
Do not bind unrelated npm packages such as `managed-agents` to this identity.

Reviewed access

Access & risk

Review the exact access boundary before connecting this MCP to accounts, credentials, or sensitive data.

Authentication: Optional Harness API key.
Permissions: Agent/session management authority of the Harness runtime.
Data handling: Session prompts, artifacts and workspace data are managed by the self-hosted API.
Cost: free
Risk and cautions: The MCP bridge can launch and control autonomous coding-agent sessions; protect the Harness API and workspace boundaries.

Access scope and cautions

Write actions: Create/run/stop agent sessions
Sensitive access: Agent workspaces Artifacts API key
Risk level: high

FAQ

Based on reviewed evidence as of 2026-10-06. Watchlist tracking is not a recommendation or a security certification.

Which tasks are documented for this implementation?

The reviewed tasks include: List agents; Create/run durable sessions; Inspect session/artifacts; Stop sessions.

What setup and authentication are documented?

Recorded installation/entrypoint step: docker run --rm -i -e MANAGED_AGENTS_URL=http://host.docker.internal:3000 ghcr.io/sandbaseai/sandbase-harness-mcp:0.3.8. Recorded transport: stdio. Recorded environment-variable notes: MANAGED_AGENTS_URL; MANAGED_AGENTS_API_KEY. No configuration values are supplied here. Prerequisites: Node.js 22+; Running Harness API; Docker for published image. These are the documented setup details, not a claim that installation or client compatibility was independently tested. Authentication: Optional runtime API key if auth is enabled.

Can it modify data or perform write operations?

Documented write actions include: Create/run/stop agent sessions. Recorded permission boundary: Agent/session management authority of the Harness runtime.

What data-handling boundary should I consider?

Session prompts, artifacts and workspace data are managed by the self-hosted API. Recorded sensitive-access context: Agent workspaces; Artifacts; API key.

What limitations and risks should I check before use?

The MCP bridge can launch and control autonomous coding-agent sessions; protect the Harness API and workspace boundaries. MCP bridge requires a running Harness API. No independent runtime, installation, or comprehensive security audit was performed.

Is this an official MCP?

SandBase AI is listed as First-party MCP in MCP100’s current evidence.